Spain Proposes AI Law with €35M Fines for Unlabelled Deepfakes, Aligning with EU Regulations
Image Credit: Florian Wehde | Splash
The Spanish government has approved a draft law aimed at regulating AI-generated content, including deepfakes, following a decision by the Council of Ministers on March 11. The legislation introduces transparency requirements for companies and addresses risks posed by manipulated media. Still pending approval from Spain’s lower house, the bill aligns with the European Union’s broader efforts to oversee AI technologies.
[Read More: Riding the AI Tide: Europe's Ambitious Leap into Next-Gen Combat Aviation]
Fines for Failing to Label AI Content
The proposed law sets significant penalties for companies that do not properly label AI-generated videos, images, or audio. Fines could reach up to €35 million or 7% of a company’s global annual turnover, depending on which amount is higher. Spanish officials have categorized such violations as a “serious offence”, reflecting concerns over the spread of deepfakes—content created or altered by AI to mislead viewers. Digital Transformation Minister Óscar López emphasized the issue’s seriousness, telling Reuters, “Everyone is susceptible to these kinds of attacks”, and pointing to potential threats to democracy.
Analysts note that while the fines aim to enforce accountability, they may not fully prevent the creation or distribution of deceptive content.
[Read More: Could a Hypothetical Deepfake Scandal on Election Eve Threaten Democracy?]
Consistency with EU Regulations
The legislation incorporates elements of the European Union’s AI Act, which is slated to take full effect across the bloc by August 2026. Spain’s move to adopt these rules early makes it one of the first EU nations to translate the bloc’s guidelines into national policy. The EU AI Act requires transparency for high-risk AI systems, a standard reflected in Spain’s mandate to label AI-generated outputs. This alignment ties Spain’s efforts to a wider European regulatory approach.
Spain’s early action suggests it could influence how other EU countries implement similar rules. The law’s consistency with EU goals also reinforces Spain’s position in regional AI governance discussions.
[Read More Understanding the Risk Classification of the EU's AI Act]
Protections Against Manipulation
The draft law includes measures beyond labelling requirements to address AI’s potential misuse. It prohibits subliminal techniques—such as subtle audio or visual cues—intended to influence behaviour without detection. Minister López highlighted examples like chatbots promoting gambling to vulnerable individuals or toys prompting children to undertake unsafe actions, illustrating the risks targeted by this rule. The provision seeks to protect those less equipped to identify manipulative tactics.
The legislation also bars organizations from using AI to assess individuals based on biometric data, behaviour, or personal traits for decisions about benefits or criminal risk. Exceptions are allowed for national security purposes, such as real-time biometric surveillance by law enforcement in public areas. Some critics question whether this exception might affect privacy rights, while others see it as a practical compromise for safety.
Enforcement by New AI Agency
Responsibility for overseeing the law falls to the Spanish Artificial Intelligence Supervisory Agency (AESIA), a newly formed entity based in A Coruña. While AESIA leads in AI oversight, other regulators handle related domains like data protection and elections, ensuring comprehensive governance across intersecting areas.
Concerns exist regarding AESIA's resources and readiness. The agency plans to expand its workforce to approximately 80 employees by the end of 2025, with current staffing including a director general, a secretary general, two division heads, a deputy director, and administrative staff. Funding is provisionally sourced from the Ministry for Digital Transformation until AESIA establishes its own budget. Its performance will be key as Spain addresses the challenges of regulating a rapidly changing field.
[Read More: UK Advances AI Governance: New Laws, Innovation Office, and Quantum Centre Unveiled]
Source: Reuters, La Moncloa
